| This essay collection explores topics pertinent to S.T.E.M education policy. The first two essays focus on stereotype threat, which posits that when individuals are primed about a fixed characteristic that is negatively stereotyped in relation to a task, subsequent task performance is adversely affected. This theory offers one factor for why test score gender gaps in math arise and women are underrepresented in S.T.E.M professions. Despite many confirmatory laboratory studies, complementary field studies of stereotype threat are notably lacking. The first essay exploits a design quirk in a nationally representative U.S. math exam, which randomly assigns students to test blocks, some of which include gender primes while others do not. I test for stereotype threat by comparing the test score gender gap for students receiving gender primes to those who do not, and find that girls perform better with some gender primes and no worse with others.;The second essay probes this apparent sensitivity of stereotype effects to how primes are worded with two analyses: a meta-analysis correlating cross-study variation in primes and stereotype findings, and a laboratory experiment manipulating the gender primes given. The results show that stereotypical primes (presenting a stereotype in the presumed direction) predict stereotype threat, except when also self-affirming (allowing students to re-affirm their rejection of the stereotype), in which case girls actually perform better. These two essays together suggest that gender primes affect test scores even in broader populations and field settings. However, how primes are framed is a key mediator of the resulting effects.;The third essay empirically explores the relationship between job search methods and turnover, satisfaction, and productivity for postdoctoral scientists. Because of the ad hoc postdoc placement process, significant variation in search methods exists for testing the social network theory that personal connections in job searches lead to superior outcomes. Using data on postdocs in the U.S., we find that foreign postdocs more often use "impersonal" searches and appear to have worse job matches, which relates to differences in outcomes. This suggests that policies facilitating job searches could yield beneficial effects, especially for the increasingly important foreign postdoc population. |
